Jump to content

Sawyer

Recommended Posts

Hi folks,

I was annoyed by running into over-voltage trouble everytime I start with a fully charged battery and have to run downhill. For some time now I used a modified charger with reduced end-voltage, but I wanted to solve the problem by its root. Together with electronics developer friends we created a circuitry that activites a load resistor, whenever a certain voltage level (67.3V) is exceeded. Such a circuitry is called break chopper. It helps breaking any kind of motor/generator drive. I installed high power LEDs (2*100W) into my KS16 as the load resistor and I am fairly satisfied with the function. I also made experiments with halogen lamps as the load resistor, what I show in the video too. Such lamps radiate away the largest part of the braking energy so it has not be spread by heat sinks.

Besides the LEDs, in the second part of the first video I use 3*150W /24V halogen lamps as the load, but these seem a little oversized.

In the second video I do the same track (140m descent) with 3*75W /24V lamps, what works as well as the 200W LED-load. 

I hope you enjoy the nerdy experiments and light show!

 

 

 

Link to comment
Share on other sites

@Sawyerthat looks a brilliant solution, with the added benefit of telling you when regen braking is active (at least when the battery is nearly full!)

i particularly like the idea of the high wattage leds inside the case giving an eerie glow beneath the wheel, kind of like the wheel going into meltdown ;-)

Link to comment
Share on other sites

This is really cool, I was researching into how to dissipate the extra voltage during braking sometime in later 2015, as I had issues with battery BMSs that cut the power due to too high voltage during braking. It never occurred to me to use LEDs or halogens for load, and I discarded the idea of using a huge braking resistor or TVSs, as they'd probably overheat :P

Thanks for the schematic too! I grabbed a screenshot of it from the video, in case someone else wants it:

OZ8khhz.png

Hope I understood this right, here's how I think it works, correct me if I'm wrong:

The +70V is the voltage to/from the motor, R1-R4 are used to set the trigger point where the load (halogens and/or leds) turns on. There's a zener-diode D1 to keep the voltage from raising too high / preventing it from going too far below 0V (negative voltage) and a smoothing (bypass) capacitor C2. This is fed into the non-inverting input of the first unit (U1A) of a 393 (a dual comparator). The inverting input comes from +5V rail through resistor R7 and is set to 4.096V with what appears to be a precision voltage reference (LM4040A411DBZT).

The second unit (U1B) is used to detect over temperature (of the leds/halogens?) with trigger voltage set from the 4.096V reference through a potentiometer (U4), the second (inverting) input of the second unit comes from an external temperature sensor, which apparently gives out a voltage signal (could be simply something like an NTC with a resistor to form a voltage divider from the +5V input?).

EDIT: @Sawyer corrected me below, the temperature sensor gives out a current-, not a voltage-signal, that's what resistors R6 & R9 are there for (they cause a varying voltage drop depending on the sensor output current, the comparator reacts to voltage).

The comparators have open-collector (or open-drain?) outputs, meaning they just start conducting to ground when they "fire". The resistors R5 and R10 are used as pullups, so the voltage coming "out" from the comparators is 5V when they're not triggered (actually, it's coming from the 5V rail through the pull-up resistor, so the "output" goes to 5V when the comparator is not conducting and pulls the voltage to ground-level when it starts to conduct). Do note that the signals and reference voltages that they are compared to are in "reverse" between the units (the reference voltage goes to non-inverting input in one and into inverting input in the other).

The outputs of them are fed to U2B (74HCT74, a dual set/reset D-flip-flop, the second unit marked as U2A is unused), which is used to control a mosfet (BUK969R), that also has pulldown from the gate to ground to make sure it isn't left floating at any point (and turns off without any signal from the flip-flop). The actual load is then turned on and off through the low-side mosfet. I didn't check the datasheet, but apparently the output of the comparator unit on the temperature sensor-side goes to the reset (marked as 2RD) to turn off of the flip-flop, to keep the flip-flop output low when there's too much heat, even if there's overvoltage. 

The flip-flop needs a "clock"-signal (2CP) to read the state from the data input (2D), that's fed from a basic 555-timer circuit.

Capacitors C1, C4 and C7 are bypass caps for the ICs, meant to be placed as near as possible to the power supply pins of the three ICs.

That's really neat, I would have personally probably skipped the whole flip-flop and just used a comparator, but the overheating protection is a very good thought (yeah, it does allow the voltage to go "too high" when it overheats, but that would happen anyway if the leds/halogens burn due to overheating, so better save the expensive parts ;)).

Btw, can the 100W LEDs take the voltage as-is (ie. 70V or such) or is something like a separate switching power supply needed to drop it somewhere?

 

Link to comment
Share on other sites

19 hours ago, esaj said:

 

Btw, can the 100W LEDs take the voltage as-is (ie. 70V or such) or is something like a separate switching power supply needed to drop it somewhere?

He answers that in German in the Video, the LEDs have a forward voltage of 33V, so two in series match perfectly with the 67,2V of the overvoltage threshold.

The rest of your post is a good summary (well, actually more an in depth view) of his explanations in the video :)

Link to comment
Share on other sites

+esaj, 

Your description is correct! The temperature sensor (not drawn) is an AD592, it is a current source, that creates a voltage across R6+R9, I recommend a capacitor (10nF) there to suppress noise. I took the FlipFlop thing for a reason. Earlier experiments showed that voltage comparators always have a linear transition region with non-definite state that does not fully drive the FET. The FET got hot an blew up. That occured still with a trigger gate inbetween. But it was a different comparator and FET too. The FF makes nice steep transitions into always a defined state. I can feel no warming when I touch the FET now.

The 4.096V reference is much more temperature stable than a z-diode, maybe important for an outdoor gadget.

I bought 6 LED for 30€ at ebay and measured the voltage necessary for 3A. (3.5A absolut max, 80°C max) They differ by a few volts and I selected two that sum up slightly below 68V. Closest data sheet I found for the LEDs belongs to LED-P100-D from WayJun Technology.

I took 5V supply from the LED-Ring of the KS16, so it deactivates when wheel is off, FET-gate is pulled down.

I recomment setting the tripping point like this: fully(est) charge the battery, leave charger connected, that is the highest ever occuring voltage in rest. Activate the circuit (5V supply). Set potentiometer so that the LED/Halogen is just not activated or flickering. By this procedure you have no hassle with accurately measuring and setting these voltages, and the wheel mainboard does have some deviation in measureing the voltage too.

Link to comment
Share on other sites

You can probably build it with fewer IC, on one of my boards I have an IC that uses a mosfet to form something similar to a relay, it has inbuilt voltage reg to drive the gate and it could be controlled by an IC that activates at certain voltages. In theory you would need trigger, gate control, LDO vsupply + passives. Of course it would be better if it were just built into the unicycle. 

Link to comment
Share on other sites

What does the driver in the car behind think? You ask.  He's thinking ? goodness gracious great balls of fire !!??

what a great project, I want one.  I frequently have to fighti lassie in her over charged state.  Goes on for over a km sometimes.

Link to comment
Share on other sites

On 12.2.2017 at 8:10 PM, Sawyer said:

Hi folks,

I was annoyed by running into over-voltage trouble everytime I start with a fully charged battery and have to run downhill. For some time now I used a modified charger with reduced end-voltage, but I wanted to solve the problem by its root. Together with electronics developer friends we created a circuitry that activites a load resistor, whenever a certain voltage level (67.3V) is exceeded. Such a circuitry is called break chopper. It helps breaking any kind of motor/generator drive. I installed high power LEDs (2*100W) into my KS16 as the load resistor and I am fairly satisfied with the function. I also made experiments with halogen lamps as the load resistor, what I show in the video too. Such lamps radiate away the largest part of the braking energy so it has not be spread by heat sinks.

Besides the LEDs, in the second part of the first video I use 3*150W /24V halogen lamps as the load, but these seem a little oversized.

In the second video I do the same track (140m descent) with 3*75W /24V lamps, what works as well as the 200W LED-load. 

I hope you enjoy the nerdy experiments and light show!

Grats! Great idea and implementation!

Some thoughts that arose reading your post:

As it seems from your video the ~200W Leds dissipate enough power going downhills with your ?KS16?.

- But if 100kg go down a 10° slope with 30 km/h i come to ~1420W of potential energy "released". (P=mg Delta h/Delta t, v=8,3 m/s, vertical Speed=Delta h/Delta t=8,3*sin(decline angle) ). From this around 160W are "dissipated" by rolling friction, about ~280W by air drag and something by the motor coils and the mosfets themselves.

That is the energy needed to lift some weight up - can this be taken as the energy released while a mass "travels" downwards? Or would some different ways be needed to calculate the power to keep the speed going down constant?

- Braking 100 kg from 30km/h within 30m should need ~1900W ( time to brake = v/2 / s, E=mv²/2, P=Delta E/Delta t )

- @zlymexposted in http://forum.electricunicycle.org/topic/6022-breaking-crash-analysis/?do=findComment&comment=73480 a real current measurement while regenerative braking with his GW. Going down moslty the current was below and around 5A (325W) with peaks up to 10 (650W) to 15A (975W) (with around 65V while reg braking in his case).

I am exitedly looking forward to further reports from you experience with this device!

On 14.2.2017 at 3:03 AM, Sawyer said:

I bought 6 LED for 30€ at ebay and measured the voltage necessary for 3A. (3.5A absolut max, 80°C max) They differ by a few volts and I selected two that sum up slightly below 68V. Closest data sheet I found for the LEDs belongs to LED-P100-D from WayJun Technology.

According to the datasheet the forward Voltage at 3,5A is between 32 and 36V. The Charateristic curve at page 3 (Forward Current over Forward Voltage) shows ~29,6V at 3,5A - going up to 4A with on the other side 3,5A given as absolute maximum rating.... The datasheet seems to be a bit inconsistent ... ;(

However as your specific LEDS have ~3A at 68/2V thats perfect - just if someone wants to rebuild this, one has to be aware that the forward voltage of such Leds could have quite some variation in forward voltage!

Regarding the overtemp shutoff of your circuit:

Don't you think some alarm (acoustic?) would be better than stopping the LEDS to not introduce the faceplant prob again? Of course then one should quite immedeately stop the wheel so the LEDs wont burn.

 

Link to comment
Share on other sites

On 19.2.2017 at 10:27 PM, Chriull said:

Grats! Great idea and implementation!

Some thoughts that arose reading your post:

As it seems from your video the ~200W Leds dissipate enough power going downhills with your ?KS16?.

- But if 100kg go down a 10° slope with 30 km/h i come to ~1420W of potential energy "released". (P=mg Delta h/Delta t, v=8,3 m/s, vertical Speed=Delta h/Delta t=8,3*sin(decline angle) ). From this around 160W are "dissipated" by rolling friction, about ~280W by air drag and something by the motor coils and the mosfets themselves.

That is the energy needed to lift some weight up - can this be taken as the energy released while a mass "travels" downwards? Or would some different ways be needed to calculate the power to keep the speed going down constant?

- Braking 100 kg from 30km/h within 30m should need ~1900W ( time to brake = v/2 / s, E=mv²/2, P=Delta E/Delta t )

- @zlymexposted in http://forum.electricunicycle.org/topic/6022-breaking-crash-analysis/?do=findComment&comment=73480 a real current measurement while regenerative braking with his GW. Going down moslty the current was below and around 5A (325W) with peaks up to 10 (650W) to 15A (975W) (with around 65V while reg braking in his case).

I am exitedly looking forward to further reports from you experience with this device!

According to the datasheet the forward Voltage at 3,5A is between 32 and 36V. The Charateristic curve at page 3 (Forward Current over Forward Voltage) shows ~29,6V at 3,5A - going up to 4A with on the other side 3,5A given as absolute maximum rating.... The datasheet seems to be a bit inconsistent ... ;(

However as your specific LEDS have ~3A at 68/2V thats perfect - just if someone wants to rebuild this, one has to be aware that the forward voltage of such Leds could have quite some variation in forward voltage!

Regarding the overtemp shutoff of your circuit:

Don't you think some alarm (acoustic?) would be better than stopping the LEDS to not introduce the faceplant prob again? Of course then one should quite immedeately stop the wheel so the LEDs wont burn.

 

Your power calculations are right. I assume that friction consumes a lot more power than 160+280W to explain why ~200W are enough to dissipate the energy from a steady roll downhill.

At the moment I don't have the tools to measure and log the real current in the leads towards the mainboard. The current displayed in the KS16 is said to be measured in the motor phases. When I found a way to record the net current, I will post the results here.

No, I think an overtemp alarm is not necessary, because a shut-off will first only lead to an overvoltage warning and tilt-back and not immediately to a face plant. The cooling fins did get only moderately warm during that try. They are exposed to the air flow next to the rim. 

Link to comment
Share on other sites

  • 6 months later...

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...